ViewVC Help
View File | Revision Log | Show Annotations | Download File
/cvs/Coro/Changes
(Generate patch)

Comparing Coro/Changes (file contents):
Revision 1.373 by root, Wed Sep 24 09:16:16 2008 UTC vs.
Revision 1.482 by root, Mon Dec 15 16:22:34 2008 UTC

1Revision history for Perl extension Coro. 1Revision history for Perl extension Coro.
2 2
3TODO: should explore PerlIO::via::CoroCede and PerlIO::via::CoroAIO. 3TODO: should explore PerlIO::coroaio (perl leaks like hell).
4TODO: maybe implement a default message channel, very much like Erlang's 4TODO: maybe implement a default message channel, very much like Erlang's
5 actor model (which is cool in a lot of important aspects (failures!), 5 actor model (which is cool in a lot of important aspects (failures!),
6 but very lacking in others (higher level ipc)). 6 but very lacking in others (higher level ipc)).
7 7
84.748 85.13
9 - EXPERIMENTAL: implement dynamic winds (on_enter/on_leave).
10 - don't set diehook to C<undef> but instead to NULL, to avoid
11 spurious warnings.
12 - fix a lot of bugs in Coro::SemaphoreSet.
13 - Coro::SemaphoreSet will less often create a semaphore needlessly.
14 - add Coro::SemaphoreSet::count and wait methods.
15 - take advantage of the new Guard module.
16 - deprecate Coro::guard.
17 - try to fix the dreaded 01_unblock tests once more. I hate it when
18 testsuites need more fixing than the code they are supposed to test.
19 - croak in more cases when a required callback isn't resolvable.
20 - fix some minor issues in Coro::State->call/eval.
21 - use current coroutine context instead of a temporary one
22 when temporarily switching to another coroutine.
23
245.12 Sun Dec 7 13:30:38 CET 2008
25 - add default config for MirOS, which seems to be a bug-to-bug
26 compatible fork of openbsd ("world domination by releasing
27 openbsd cvs before the openbsd folks do it" or so :).
28 - free_padlist did destroy the names pad, not good, but didn't
29 seem to bother perl - this could fix issues such as eval ""
30 inside a function called from multiple coroutines.
31 - use a different method to detect destruction time.
32 - be more careful when freeing padlists just before global
33 destruction.
34 - fixed and expanded the call/cc example.
35 - renamed _terminate to _coro_run.
36 - new method Coro::Channel->shutdown.
37 - try pthreads on openbsd <4.4 (broken sigaltstack, will
38 pthreads fare better?).
39 - be less picky about destroying "a" running coroutine.
40
415.11 Tue Nov 25 21:49:05 CET 2008
42 - DEBUGGING in 5.10.0 is a mess: it constantly flags perfectly
43 working code with failed assertions, introducing more bugs than
44 it fixes, requiring elaborate workarounds :(
45
465.1 Mon Nov 24 08:54:59 CET 2008
47 - wrote a small introductory tutorial - Coro::Intro.
48 - convert Coro::Timer, Coro::Select and Coro::Util to rouse API.
49 - Coro::Select did errornously dup the file descriptors
50 and didn't work with all AnyEvent backends.
51 - Coro::Select wasn't imported correctly form Coro::LWP, causing blocking
52 LWP data transfers.
53 - disassociate c contexts from coro objects - this is agruably more
54 correct, but mostly allows sharing of cctxs between coro and state
55 objects, for added memory savings and speed increases.
56 - bumped $Coro::POOL_RSS up to 32kb by default.
57 - no longer set the optype to OP_CUSTOM, as B::* understandably
58 doesn't like this very much (and we *are* a type of entersub).
59 - implement state cloning, just to prove that call/cc can be done.
60 - automatically load Coro::AnyEvent in Coro::Handle.
61 - wrap ->cancel calls in eval inside Coro::Handle as EV watchers
62 do not have this method (and don't need it either).
63 - speed up generic anyevent methods in Coro::Handle by using rouse
64 callbacks.
65 - allow coroutines in $Coro::IDLE, speeding up Coro::AnyEvent and
66 others. It also makes the debugger happier, as you can trace
67 through the idle threads now.
68 - add comppad_name* and hints ($^H, %^H) to per-thread variables.
69 - eg/event was pretty broken.
70 - better 5.8.6 compatibility.
71
725.0 Thu Nov 20 10:35:05 CET 2008
73 - NEW ARCHITECTURE: use the latest 4.x version if you experience
74 stability issues.
75 - bump API version to 7 - all dependents must be recompiled.
76 - removed timed_* functions - they were not being used anyways
77 and should be replaced by a more generic mechanism -
78 and were annoying to support anyways :)
79 - removed SemaphoreSet's waiter method - use sem method instead.
80 - Coro::Semaphore->adjust didn't correctly wake up enough waiters.
81 - async_pool did free a scalar value twice
82 ("Attempt to unreference...").
83 - fix a longstanding bug where calling terminate on a coro that
84 was waiting for a semaphore that was just becoming available
85 would cause a deadlock (semaphore would get into a state where
86 it was available but waiters were still blocked).
87 - calling throw on a coroutine that is waiting for a semaphore will
88 no longer make it acquire the semaphore (and thus leak a count).
89 - perl's process emulation is now not even theoretically supported
90 anymore.
91 - new functions Coro::rouse_cb and Coro::rouse_wait for easier
92 conversion of callback-style to blocking-style.
93 - new methods $coro->schedule_to and ->cede_to, to specifically
94 schedule or cede to a specific coroutine.
95 - new function Coro::Semaphore::wait.
96 - use named constants in Coro::Channel (Richard Hundt).
97 - directly patch the entersub opcode calling SLF functions (cede,
98 transfer and so on). this does speed up context switching, but
99 more importanly, it frees us from the hardcoded behaviour of
100 entersub, so we might actually be able to return something from
101 those functions and atcually create new ones.
102 - take advantage of __builtin_frame_address on gcc.
103 - expose THX in coroapi (not sure whether this was a wise decision,
104 as "threaded" perls are running at half speed anyways).
105 - implement execute_slf (schedule-like-function) interface that makes
106 it possible to implement schedule-like-functions in XS.
107 - use new SLF interface to massively speed up Coro::EV by roughly a
108 factor of two.
109 - used new SLF interface to massively speed up Coro::Semaphore by a
110 factor of three.
111 - used new SLF interface to speed up Coro::AIO by roughly a factor of
112 four and reduce its memory usage considerably.
113 - implement Coro::SemaphoreSet purely in terms of Coro::Semaphore,
114 for a nice speedup and vastly more correct behaviour. Also implement
115 a new method "sem" to get at the underlying semaphore object.
116 - implement Coro::Channel in terms of Coro::Semaphore, for a moderate
117 (in comparison) 20-40% speedup.
118 - used new SLF interface to reimplement Coro::Signal gaining
119 some unknown (because I was too lazy), but certain, speedup, and also
120 making signals reliable for the first time.
121 - used new SLF interface and other optimisations to speed up async_pool
122 by a factor of two. It also doesn't rely on perl's exception mechanism
123 to exit anymore. The overhead for terminating an async_pool, coro over
124 a normal async is now very small.
125 - sped up coroutine creation/destruction by 40%.
126 - forgot to include Coro/libcoro/README in the dist for all these years.
127 - work around a freebsd pthreads bug (manual testcancel is required as
128 pthread_cond_wait isn't a cancellation point on freebsd).
129 - use new rouse functions to speed up and simplify Coro::BDB.
130 - make "prefer perl native functions" work with threaded perls.
131 - condense Coro::Debug ps output, hint at v and w flags.
132 - (libcoro) lots of minor cleanups and portability improvements.
133
1344.914 Wed Nov 19 12:54:18 CET 2008
135 - fix a disastrous bug in the readline optimisation
136 introduced in 4.801.
137
1384.913 Sat Nov 15 07:58:28 CET 2008
139 - async_pool did free a scalar value twice
140 ("Attempt to unreference...").
141
1424.912 Thu Nov 13 18:31:23 CET 2008
143 - minor cleanups.
144 - use much larger stacks on linux and perl < 5.8.8.
145 - Coro::Debug::new_unix_server did not unlink the socket
146 when destroyed.
147
1484.911 Tue Nov 11 04:26:17 CET 2008
149 - "port" to threaded perls.
150
1514.91 Mon Nov 10 05:36:38 CET 2008
152 - the ->throw exception object no longer leaks.
153 - creating a new cctx leaked a scopestack entry (memleak).
154 - new coroutines didn't get created with a zero flags field
155 (unknown impact).
156 - calling ->throw on a not-yet-started coroutine should now work
157 instead of being ignored.
158 - ->throw is now supported on Coro::State objects.
159 - clean up cctx creation code a bit.
160 - entersub is actually a UNOP, not a LOGOP (not a bugfix).
161
1624.9 Sat Nov 8 17:45:27 CET 2008
163 - (libcoro) did not preserve rbp with CORO_ASM (we are getting there).
164 - (libcoro) no longer leak threads in the experimental pthread backend,
165 also speed it up considerably.
166 - (libcoro) do not rely on makecontext passing void *'s unscathed.
167 - fix compiletime dependencies on libcoro in the Makefile.
168 - cctx_count wasn't always updated properly.
169 - Coro::State::cctx_stacksize wasn't applied correctly.
170 - new function Coro::State::cctx_max_idle.
171 - the default max number of idle C contexts is now 4.
172 - (libcoro) try harder to get _setjmp/_longjmp.
173 - (libcoro) cleanup and extend the libcoro API to officially
174 allow the creation of empty source contexts.
175 - very experimental workaround for the totally broken netbsd platform.
176 - tried to hack around openbsd bugs.
177
1784.804 Wed Nov 5 16:36:00 CET 2008
179 - Coro::Debug::new_unix_server would not create a non-blocking listening
180 socket, sometimes causing freezes.
181 - (libcoro) fix misaligned stack points for setjmp and assembly
182 methods, which can cause crashes on x86/x86_64 with a sufficiently
183 aggressive compiler.
184 - new function: Coro::Debug::new_tcp_server.
185 - move ->throw into the Coro class because it only works on coro objects.
186
1874.803 Mon Nov 3 17:16:12 CET 2008
188 - (libcoro) use a global asm statement to become independent of gcc
189 otpimisations for CORO_ASM (thanks to pippijn for the idea).
190 - try to workaround yet another broken bsd, this time dragonfly.
191
1924.802 Thu Oct 30 10:56:12 CET 2008
193 - support -fno-omit-frame-pointer on x86 with the assembly method.
194 - tune 01_unblock.t tests a bit.
195
1964.801 Wed Oct 22 18:33:37 CEST 2008
197 - improve readline speed for very long "lines".
198 - backport to 5.8.8.
199
2004.8 Thu Oct 2 13:34:40 CEST 2008
201 - new function Coro::AIO::aio_wait.
202 - Coro.:AIO and Coro::BDB now "use Coro::AnyEvent".
203 - greatly speed up and reduce memory usage of Coro::AIO requests.
204 - implement some other ยต-optimisations.
205
2064.749 Mon Sep 29 14:40:12 CEST 2008
207 - port to slow and broken pseudo-threaded perls. (courtesy pippijn).
208
2094.748 Sat Sep 27 14:03:19 CEST 2008
210 - implement, but do not document, PerlIO::cede(granularity).
9 - Coro::Storable forgot to wrap Storable::pstore. 211 - Coro::Storable forgot to wrap Storable::pstore.
10 - work around a memory leak in perl in Coro::Storable 212 - work around the multitude of leaks and memory corruption
11 (and note that perl 5.10.0 has a double-free bug 213 bugs in PerlIO::via by using our own C-level perliol. As a side
12 that is not being workarounded). 214 effect, Coro::Storable is now much, much, much faster.
13 215
144.747 Tue Sep 23 01:59:41 CEST 2008 2164.747 Tue Sep 23 01:59:41 CEST 2008
15 - fix a per-cv memleak (one empty array was leaked per 217 - fix a per-cv memleak (one empty array was leaked per
16 code reference). 218 code reference).
17 - avoid a crash in coro->call|rss when the coroutine was already 219 - avoid a crash in coro->call|rss when the coroutine was already

Diff Legend

– Removed lines
+ Added lines
< Changed lines
> Changed lines